Cutaneous Drug Eruptions

Introduction

Cutaneous drug reactions account for a large proportion of adverse drug reactions. Cutaneous drug reactions can be very challenging to diagnose. They can mimic many other skin diseases; this is especially evident during childhood when viral exanthems are commonplace. If a patient is taking numerous medications, establishing causality to a specific drug can be difficult.

This chapter includes a review of general principles, mechanisms, and clinical manifestations of cutaneous drug eruptions. We have classified different types of drug eruptions by morphology: exanthematous, urticarial, pustular, and bullous. Within each of these groups we have divided them into simple, benign, or nonfebrile and complex or febrile reactions. We also include a miscellaneous group to ensure a methodical review. Diagnostic maneuvers are discussed, and an algorithm is presented to enable the clinician to attain an idea about the possible responsible drug.

There are a number of ways to classify drug reactions. Our classification is a widely accepted and simplified approach based on morphology and subdivided into simple and complex to take into account systemic features. Alternative common methods of classification are either to divide reactions into Type A (predictable, acute, related to mechanism of action) or Type B (idiosyncratic, unpredictable, and not related to mechanism of action), versus classifying reactions based on the type of immunologic response (immediate IgE-mediated hypersensitivity reactions, cytotoxic reactions, immune-complex-mediated reactions, and T-cell-mediated/delayed type reactions).

Epidemiology of Cutaneous Drug Reactions

Cutaneous adverse drug reactions (CADRs) are a commonly reported type of adverse drug reaction (ADR) and can lead to frequent clinical visits and discontinuation of therapy. Dermatologic reactions are the most common manifestation of systemic drug hypersensitivity. Up to 2% to 3% of all hospitalized patients experience either an urticarial or an exanthematous drug eruption. Up to 5% of patients who receive certain antibiotics while hospitalized experience either urticarial or exanthematous reactions. During hospital admission 10% to 20% of patients have a drug reaction, and they represent the fifth most common cause of death in hospital.

When a combination of trimethoprim–sulfamethoxazole is given to human immunodeficiency virus (HIV)-infected patients a hypersensitivity reaction may develop in as many as 50%. The risk of severe drug eruptions such as Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) is increased in HIV-positive patients and also in SLE, allo/BMT patients, and other immune dysregulated/immune abnormal states. Fatal anaphylaxis from intramuscular penicillin and fatal anaphylactoid reactions from radiocontrast each occur in about 1:50,000 exposed patients. In hospitalized children, cutaneous eruptions are the most common type of drug reaction seen. Around 2.5% of children receiving medication in the outpatient setting will experience a drug eruption, and this figure rises to 12% if the drug is an antibiotic.

Drug-Induced Skin Injury

The International Serious Adverse Event Consortium in collaboration with other stakeholders initiated the Phenotype Standardization Project. The goal was to develop standardized phenotypic definitions for three types of ADRs including cutaneous reactions caused by drugs: “drug-induced skin injury” (DISI). While the majority of DISIs are mild, the more serious entities including drug-induced hypersensitivity syndrome (DIHS), which is also known as drug reaction with eosinophilia and systemic symptoms (DRESS), SJS/TEN, and acute generalized exanthematous pustulosis (AGEP) form a more complex scenario within which genetics are an emerging factor.

Approach to the Patient with a Suspected Drug Eruption

Making the correct diagnosis is a vital element in the assessment process of possible CADRs prior to introducing treatment or recommendations. Diagnosis can be difficult because drug eruptions can closely mimic other diseases (e.g., exanthematous drug reaction vs. a viral exanthem, toxin-mediated erythema, or acute graft-versus-host disease, and AGEP vs. pustular psoriasis). Identification of the causative drug can be complicated if the patient is taking several different drugs concomitantly. For accurate diagnosis a rational approach is required. This includes an initial clinical impression, forming a differential diagnosis, analysis of drug exposure (including timing of new medications, dose adjustments or increases, drug–drug interactions, and metabolic changes and their impacts on drug levels, such as renal or hepatic insufficiency), laboratory results, diagnostic tests, and utilization of the available literature. Prioritization of the diagnosis includes a causality assessment.

An initial clinical impression is based principally on the morphology of the eruption. The four main categories described, based on the primary lesion, are exanthematous, urticarial, blistering, or pustular. Systemic signs (e.g., malaise, fever, hypotension, tachycardia, lymphadenopathy, synovitis, dyspnea, etc.) allow for refinement of the primary clinical impression ( Table 47-1 ). These systemic signs may aid in differentiating a benign or simple cutaneous drug eruption from a severe or complex drug eruption ( Table 47-2 ). Establishment of the diagnosis is the final step in diagnosis. Table 47-3 identifies the target organs potentially involved in complex reactions.

A careful analysis of drug exposure should ensue. All medications should be included, regardless of route of administration, and incorporate prescription drugs, over-the-counter, and herbal remedies. Patients or caregivers should be asked specifically about vitamins, pain medications, sedatives, laxatives, oral contraceptive pills, and any medications that may not be on a pharmacy list (e.g., infliximab, radiocontrast dye). In the Boston collaborative drug study most cutaneous reactions occurred after the first week of exposure. The onset of the cutaneous reaction should be carefully documented as well as the effect of previous patient-initiated rechallenge and dechallenge. Each drug, its dose, and duration along with relevant signs and symptoms should also be recorded.

A literature search may provide helpful information. Laboratory tests can aid a diagnosis. A complete blood count and differential, liver and renal function should be requested. Cutaneous biopsies can be useful to differentiate between differential diagnoses but biopsies do not allow for confirmation of a causative drug. Histopathologic findings may help clarify the drug reaction pattern, but they do not identify the responsible drug. Histopathologic examination can confirm the diagnosis of SJS, fixed drug eruption, vasculitis, and erythroderma, and may support the clinical diagnosis of urticarial or morbilliform drug reactions. Eosinophils are widely believed to be major participants in many cutaneous drug reactions. The microscopic presence of eosinophils certainly suggests a drug cause; however, the absence of eosinophils does not exclude a drug as a possible etiologic agent nor does the presence of eosinophils confirm a drug as a possible etiologic agent.

There is no single diagnostic test that can be employed across the board in cases of cutaneous drug hypersensitivity. This is because of the variability of pathogenetic mechanisms operating in the different morphologic variants, the possibility that drug–virus interactions were important clinically, or that nonpharmacologic additives or excipients were responsible.

A list of available in vitro and in vivo diagnostic tests for drug hypersensitivity is found in Table 47-4 . In vitro testing includes lymphocyte transformation test, lymphocyte toxicity assay, histamine release test, basophil degranulation test, passive hemagglutination lymphocyte transformation test, leukocyte and macrophage migration inhibition factor tests, and radioallergosorbent test. Specific IgE assays, such as the radioallergosorbent test, are the most commonly employed for evaluating immediate hypersensitivity reactions. These include urticaria, angioedema, and anaphylaxis. Only a few drugs can be tested this way, such as the β-lactams and insulin. Although IgE assays are still less sensitive than scratch tests, they should be used together with scratch testing under proper supervision in patients at risk for anaphylaxis. The basophil activation test uses flow cytometry to detect markers of response to drug allergens. It has been employed in cases of immediate hypersensitivity to β-lactams, muscle relaxants, and nonsteroidal anti-inflammatory drugs (NSAIDs).

The lymphocyte transformation test measures the proliferative response of a patient’s T cells in vitro to a suspected drug culprit. It has been reported to be more sensitive for diagnosis than patch testing, but has some limitations. First, although it has been found to be positive in the majority of cases of exanthematous reactions, the drug hypersensitivity syndrome, and AGEP, it is only rarely positive in cases of TEN, fixed drug eruption, and vasculitis. Second, timing of the test is important, with cases of the drug hypersensitivity syndrome showing a negative test in the first few weeks after onset of the eruption. Finally, the test is not available at most clinical centers.

In vivo tests include skin testing and provocation or oral rechallenge and also patch testing. Prick tests have been found to be a useful diagnostic tool in cases of sensitivity to β-lactams and muscle relaxants used in anesthesia. Intradermal testing can be performed when prick tests are negative. To date, penicillin is the most widely used systemic drug for which intradermal skin testing is significantly reliable. Patients with the majority of important drug reactions, including SJS and TEN, exanthematous reactions, vasculitis, and erythroderma, should not undergo this form of testing.

Patch testing for cutaneous drug reactions has been studied the most vigorously of all the skin tests. Sensitivity varies depending on the type of reaction, the putative drug, the concentration of drug tested, and, for fixed drug eruptions, the site at which the patch is placed. Positive results have been obtained in cases of exanthematous reactions, fixed drug eruption, AGEP, and the drug hypersensitivity syndrome. Sensitivity has varied between 30% and 50%. The specificity and negative predictive value have not been determined.

Establishment of the final diagnosis is the final step in the diagnosis. If a definite diagnosis is not possible then a prioritization diagnosis must be completed by combining the information gathered. The traditional approach of highly probable, probable, possible, unlikely, and almost excluded are helpful. The Naranjo assessment classifies a drug reaction as definite, probable, and possible. To fulfill the criteria for a definite reaction four components must be met. These include (1) temporal relationship, (2) a recognized response to the suspected drug, (3) improvement after drug withdrawal, and (4) reaction on rechallenge. A probable reaction includes parts 1 to 3 but does not include rechallenge and a possible reaction only requires a temporal relationship.

Discontinuation of drug therapy and resumption of therapy with the drug in question at a later time may allow immunologic effector mechanisms to “recharge” fully, making large-scale discontinuation of all drugs the patient is receiving worthy of careful scrutiny. The potential for the disease being treated to worsen after drug discontinuation has to be considered in dechallenge decisions. Each case should be handled individually, with a consideration of the risks and benefits of discontinuing drug therapy. In managing patients with high-risk reaction patterns, rechallenge with the drug in question should be carried out only in very rare circumstances, when the need to know the responsible drug exceeds the risk of a severe reaction with the rechallenge. Intentional rechallenge can be performed only when a clinical presentation meets the criteria in Table 47-5 . Reports of patients with accidental rechallenge provide useful information on drug causes, but it is essential to avoid such accidental rechallenge. It is important to note that rechallenge is not optimally sensitive or specific. Despite these limitations, rechallenge, when indicated, is the best way to identify accurately the causative drug in the clinical setting. The presence or absence of drug–drug and drug–virus interactions should always be considered in this diagnostic step.

The technique of “reverse challenge” seems most reasonable and practical when there is one drug of high suspicion and several others of lower suspicion that were started simultaneously prior to the cutaneous drug eruption. The failure to reproduce the reaction when the patient receives the low-suspicion drugs increases the likelihood that the high-suspicion drug (which is not rechallenged) is responsible for the drug reaction. This method essentially clears from responsibility the drugs that were actually rechallenged.

A negative result with oral rechallenge can mean that the drug tested was not responsible for the reaction; that it was perhaps administered at too low a dose; or that the rechallenge did not reproduce all the clinical conditions for the prior cutaneous drug eruption. A positive rechallenge can be regarded with reasonable certainty as indicating that the drug tested was responsible for the cutaneous drug reaction. Again, rechallenge is not endorsed for high-risk drug reaction patterns, except in the most exceptional circumstances.

In order to exclude (to a reasonable degree of certainty) nondrug causes for the reaction pattern present the clinician should use appropriate historic and physical examination findings, along with well-directed laboratory tests. Most commonly, a variety of infectious agents can mimic the majority of cutaneous drug eruptions discussed.

Mechanisms of Cutaneous Drug Eruptions

The patient’s genetic background may be significant. Human leukocyte antigen (HLA) molecules play an important role in drug reactions, as they present antigen to T cells. Specific HLA genotypes have been shown to confer a greater susceptibility to various drug eruptions, e.g., HLA B ^∗ 1502 and carbamazepine-induced SJS, HLA B ^∗ 5701, and abacavir-induced hypersensitivity syndrome. Defective detoxification of reactive metabolites (with anticonvulsants, by epoxide hydroxylases) is thought to be responsible for a familial predisposition to aromatic anticonvulsants and sulfonamides hypersensitivity syndrome. In drug-induced lupus the acetylator phenotype is important: slow acetylators have a higher risk. Key information sources on specific cutaneous eruptions and responsible drugs are listed in Table 47-6 .

Most drugs that induce cutaneous drug eruptions have a molecular weight of less than 1000 Da, therefore they must serve as haptens for an immunologic response. A cell-based or soluble carrier protein is necessary for a drug of this size to become a complete antigen. In most instances, drug metabolites, and not the parent drug, induce the immunologic hypersensitivity. Most allergic (immunologic hypersensitivity) drug reactions should demonstrate the following features: (1) they occur in a small percentage of patients; (2) there is a history of prior exposure to the drug or a chemically related compound; and (3) there was a latency of 1 to 2 weeks between the initial exposure and the onset of the reaction and a latency of 1 to 2 days with rechallenge. Allergic drug reactions are not dose-dependent. The reaction differs from the drug’s pharmacologic effects and from other established signs of drug intolerance. The eruption should resolve with dechallenge and reappear after rechallenge with the drug in question.

Cutaneous drug eruptions that have no specific sensitization to a drug hapten are known as “pseudoallergic” or anaphylactoid reactions. Drugs such as opiates and radiocontrast material directly degranulate mast cells without prior specific antigen sensitization. Aspirin and NSAIDs may induce urticaria by effects on the arachidonic acid pathway, leading to nonspecific mast cell degranulation. Idiosyncratic reactions can lead to either organ-specific (such as the skin) or generalized hypersensitivity.

Reactivation of viruses has been observed in drug reactions, especially in the drug hypersensitivity syndrome. Whether the reactivated virus further stimulates the immune system, leading to a more severe clinical course, or whether the virus is an innocent bystander that is reactivated by drug-induced immune stimulation, is controversial.

The most common mode of drug administration leading to sensitization is topical exposure. Oral exposure leads to specific sensitization more commonly than does parenteral (intramuscular or intravenous) exposure. After specific sensitization has occurred, rechallenge by parenteral routes is significantly more risky than by oral administration. Topical exposure presents the least risk of serious reactions with rechallenge.

Cross-reactions between chemically related drug groups are important to consider when assessing cutaneous drug reactions. Most notable are the many potential cross-reactions between drugs with a β-lactam nucleus, such as the original penicillins, aminopenicillins, semisynthetic penicillins, and probably cephalosporins; this holds true for certain groups of anticonvulsant medications as well. After the patient is sensitized to one member of this broad group of drugs, other related drugs should be considered to have a potential for cross-reaction. Aspirin and the various NSAIDs may cross-react, usually by nonallergic mechanisms. Cross-reactivity between antibacterial and nonantibacterial sulfonamides, on the other hand, is extremely unlikely based on their divergent chemical structures.

Morphologic Subtypes

Exanthematous Eruptions

Simple Exanthematous Eruptions

Exanthematous drug eruptions (synonyms: morbilliform, maculopapular, or scarlatiniform eruptions) are the most common drug-induced eruptions ( Fig. 47-1 ). They occur in 1% to 5% of first-time users of most drugs. This type of drug reaction is increased in the presence of viral infections, e.g., a near 100% incidence of exanthematous reaction in patients who have Epstein–Barr virus taking penicillin. Patients with human immunodeficiency virus infections or bone marrow transplant are at increased risk. The most common classes of drugs implicated include penicillins, sulfonamides, cephalosporins, and antiepileptic medications. Exanthematous drug eruptions are characterized by erythematous macules and/or papules, usually beginning 7 to 14 days after the initiation of a new medication, and sometimes after drug discontinuation. The eruption is usually symmetrical, beginning on the trunk, becoming generalized and is without blistering or pustulation. Mucous membranes are usually spared, and facial involvement is uncommon, but palms and soles are often involved. Pruritus is the main symptom.

Pathology is nonspecific and consists of eosinophils, a mild perivascular lymphocytic infiltrate, and associated necrotic keratinocytes at the basal layer. The differential diagnosis of exanthematous rashes is very broad. Viral exanthems tend to be indistinguishable from exanthematous drug eruptions and are more common in the pediatric population. Symptoms and a comprehensive history that includes timelines are very important in helping to establish the diagnosis. Toxic shock syndrome, scarlet fever, acute graft-versus-host disease, Kawasaki disease, and juvenile idiopathic arthritis should be excluded on the basis of clinical features. A number of tests can be performed to further evaluate the patient. These include laboratory tests to evaluate internal organ involvement and rapid strep test/throat bacterial culture. Skin biopsy is generally not useful in this setting. To further complicate issues exanthematous rashes can be exacerbated by concomitant viral infections.

The eruption is self-limited, therefore the management is largely supportive. A decision whether to discontinue the implicated drug must be made. This is based on the availability of an unrelated substitute, or if the drug is of paramount importance a decision can be made to continue it and offer symptomatic treatment. The risk:benefit ratio of this option has to be carefully weighed and the evolution of the eruption meticulously monitored. Whether continuation of a drug can lead to SJS is debatable. Oral antihistamines, bland emollients, and topical corticosteroids can be used to treat pruritus. The eruption often turns to a brownish-red and fades within 7 to 14 days of discontinuation of the offending drug. Scaling or desquamation may follow. Rechallenge may lead to the reaction appearing within a few days.

Complex Eruption: Drug-Induced Hypersensitivity Syndrome

Drug-induced hypersensitivity syndrome (DIHS, also referred to as DRESS) should be suspected when an exanthematous drug reaction occurs with associated fever and internal organ involvement ( Fig. 47-2 ). DIHS has cutaneous, hematologic, and internal organ manifestations, is severe, and leads to mortality in up to 10% of individuals. It occurs usually on first exposure to the offending drug with first symptoms 2 to 8 weeks after exposure. The reaction occurs in approximately 1:3000 exposures. The most commonly associated drugs are the aromatic anticonvulsants including phenytoin, carbamazepine, and phenobarbital. Other offending drugs are lamotrigine, sulfonamides, antibiotics, dapsone, minocycline, allopurinol, and nevirapine.

Fever and malaise are usually the first symptoms and can be accompanied by cervical lymphadenopathy and pharnygitis. While cutaneous eruption associated with DIHS can be mild it is more often extensive and severe. It often begins on the face, frequently periorbitally, initially with edema and subsequently erythema and pruritus. It then spreads caudally. Facial edema and lymphadenopathy are common; hand edema has been reported in one-third of patients. The lack of mucosal involvement is a useful distinguishing feature from SJS. Reports of DIHS in the literature describe many different morphologies and include exanthematous eruptions, purpura, cheilitis, vesicles, bullae, and targets.

Atypical lymphocytosis and/or eosinophilia are typically seen early in the course. Regarding visceral involvement, the liver is the most commonly involved internal organ (about 50%); hepatitis may be fulminant and could necessitate liver transplantation. Lymphadenopathy, joint pain, and inflammation of the kidneys, central nervous system, heart and lungs have all frequently been described. Cardiac inflammation can be either acute or delayed, and concerning symptoms should prompt immediate organ reassessment. Thyroiditis can occur but is usually not noticed for 2 to 3 months after onset; other forms of delayed autoimmunity can occur following this eruption, including diabetes, vitiligo, and lupus-like syndromes. The eruption persists for weeks to months after withdrawal. During the recovery period an initial period of improvement may be followed by flare of the cutaneous and visceral manifestations. Rechallenge with the offending drug leads to reactivation of fever and erythroderma within hours. Anticonvulsants metabolized by the cytochrome P450 system can cross-react. A patient who reacts to phenobarbital, phenytoin, or carbamazepine should avoid all three medications.

Patients with DIHS should have a battery of laboratory tests to consider visceral involvement including thyroid testing, which should be repeated at 2- to 3-month intervals. In the management of DIHS prompt withdrawal of the offending drug is vital. The role of systemic corticosteroids is controversial but generally in patients with visceral involvement or severe symptoms, treatment with prednisone (1 to 2 mg/kg per day) is usually ensued; patients often require months of systemic corticosteroid treatment (50 days on average). Antihistamines and topical corticosteroids have also been used to alleviate symptoms. Intravenous immunoglobulin (IVIG) has also been used in the management of DIHS and there are some reports also of the use of cyclosporine. First-degree relatives have a higher risk of developing the same drug reactions, so counseling of family members should be considered.